Hierarchically porous membranes for lithium rechargeable batteries: Recent progress and opportunities

Abstract

Hierarchically porous membranes offer an effective platform for facilitating mass transport and ion diffusion in energy storage systems and have the potential to achieve novel battery configurations. As the vital roles such as electrodes, interlayers, separators, and electrolytes in the battery systems, regulating the membrane porous structures and selecting appropriate membrane materials are significant for realizing high energy density, excellent rate capability, and long cycling stability of lithium rechargeable batteries (LRBs). In this review, we highlight recent progress on tunable synthesis of various porous membranes for LRB applications, and discuss how the membranes with hierarchically porous frameworks or ordered channels can be employed as electrodes/separators/interlayers for improved ion/electrolyte transport and charge transfer. We also present in‐depth discussion of the structure–property‐performance relationships of the membranes based on fundamental thermodynamic and kinetic aspects in membrane formation, transport mechanism, and cell configurations. Finally, the prospects of optimizing membrane development for advanced battery applications are discussed.

INTRODUCTION

Increasing population and depleting energy sources are significant challenges to the global energy supply. The prospering smart grid systems and electric vehicles pose new demands and provide opportunities for developing renewable energy storage technology. Secondary battery systems can directly store and release energy via conversion between chemical and electrical energies.^1,2 In secondary battery systems, the rechargeable lithium metal-based batteries have attracted attention, owing to their large capacity, high energy density, and long cycling life.³

Generally, lithium metal-based batteries are composed of a cathode, anode, separator, and electrolyte. During charging and discharging process, the Li⁺ ions perform intercalation and de-intercalation between the electrodes through electrolytes.^4,5 To satisfy the globally increasing demands of sustainable energy sources and achieve high energy density LRBs, it is essential to design battery configurations with smart-material architecture and optimize their material characteristics.^6,7 Particularly, mass and charge transport, electronic and ionic conductivities, and electrochemical reaction kinetics are pivotal factors that need to be further explored to achieve high-performance LRBs.⁸

Engineered membranes for better mass transport and ion diffusion have been widely investigated, and can be applied in biomedical,^9,10 energy storage,^11–13 and filtration fields¹⁴ and so on. Generally, it is desirable to fabricate the membranes with specific characteristics, such as strong mechanical strength, high thermal stability, large porosity, uniform pore distribution, as well as high permeability, which can overcome technical and scientific challenges. Membrane technology has been used in LRBs for its unique advantages, such as flexible construction, prospective commercialization, and large-scale production. Moreover, the membranes can serve as separators in conventional battery systems, as well as electrodes and electrolytes in advancing research. Regulating the membrane structure and selecting appropriate membrane materials are significant for realizing a high energy density, excellent rate capability, and safety of LRBs. Therefore, it is important to optimize the membrane design in LRBs and further explore the preparation-structure-performance relationship of membranes.

In this review, we focus on the recent progress in membranes for LRB applications, based on the integrated design of batteries. We highlight the highly tunable synthesis of various porous membranes for LRB applications, involving key synthetic elements such as polymer, solvent/nonsolvent, and functional additives and their determination on pore construction in membrane and the mechanism, and discuss how the porous membranes are employed as electrodes/separators/interlayers for improved mass transportation and charge transfer. The structure–property relationships of the membrane based on fundamental thermodynamic and kinetic aspects in membrane formation, transport mechanism, and cell construction engineering are also concerned. In addition, the potential opportunities and challenges of membranes in emerging LRBs are discussed. As shown in Scheme 1, the main membrane preparation technologies with their characteristics for applications in LRBs are summarized.

SCHEME 1. The fabrication methods and key features of hierarchically porous membranes and their applications in LRBs

MEMBRANE PREPARATION METHODS

Selecting a suitable method for polymer membrane fabrication is crucial for obtaining the desired porous structure. In this section, we introduce the important techniques used for porous membrane preparation, such as phase inversion, electro-spinning, and 3D-printing. Based on the porous structures and chemical stability, the membranes can be used as the critical parts of electrodes/separators/interlayers in cells and the solid electrolyte materials.

Phase separation method

Phase inversion or phase separation is one of the most popular techniques for preparation of porous membranes used in LRBs.^15–17 Generally, phase inversion occurs when the equilibrium of a homogeneous polymer solution is broken, resulting in the minimization of free energy and separation of the solution into two phases.^18,19 The rate difference of mass transfer during the phase inversion generates polymer-poor and polymer-rich phases. Subsequently, the scattered polymer-poor phases form various voids, and polymer-rich phases construct the membrane matrix.²⁰

A phase diagram is generally used to demonstrate thermodynamic states. The bimodal curve is an important component of a phase diagram, and represents the boundary of liquid–liquid demixing. The bimodal curve can be determined by measuring the cloud points and calculating through the Flory–Huggins theory.^21–23 The spinodal curve can be calculated based on the bimodal curve. As shown in Figure 1, the bimodal curve can divide the phase diagram into homogeneous and unstable regions; the phase separation occurs in the unstable region, owing to the spinodal decomposition (SD) as well as nucleation and growth (NG) mechanisms. The NG mechanism generally induces cellular pore structures, while the SD mechanism leads to the formation of pore structures with bi-continuous or droplet morphology.^15,24–28

FIGURE 1. Schematic of phase diagram and pore structure

The produced membrane structure depends on the thermodynamics and kinetics of membrane formation.²⁹ The thermodynamic aspect refers to the phase equilibrium in the polymer solution component systems, while the kinetic aspect is defined as the mutual diffusion and convection of components. There are numerous controllable factors to obtain various membrane structures. Among these, polymer concentration is of vital importance, and is associated with the thermodynamic and kinetic aspects. Pore volume is closely related to the transport kinetics during phase inversion. The delayed separation may be attributed to the high viscosity of the casting solution, high polymer concentration, slow diffusion, low miscibility of non-solvent with solvent, and prolonged evaporation time, which produces macroviods. Furthermore, the solvent type, organic/inorganic additives, air humidity, compatibility of the polymer and solvent, and temperature also affects the membrane morphology.^30–33

Physical methods to modify the stability of a polymer solution include mass exchange, temperature variation, and solvent evaporation. Based on the separation approaches, it can be divided into nonsolvent-induced phase separation (NIPS), thermally-induced phase separation (TIPS), and solvent evaporation-induced phase separation (SIPS). SIPS is rarely used to prepare the porous membranes in LRBs. Hence, we focus on the application of NIPS and TIPS in the production of membranes with various morphologies.^34,35 Our group realized the hierarchical pore structure control by adjusting the casting solution with various organic/inorganic additives and polymer-additive ratio, as shown in Figure 2. The additives affect the exchange rate of the solvent and non-solvent, which further influences the membrane structure. Rather than the macropores in SiO₂/CeO₂ doped PAN membranes and the corresponding carbon membranes (Figure 2B,C), more regular and longer finger-like pores can be observed in Fe³⁺-doped carbon membranes (Figure 2D,E). This structure results from the addition of metal salt, which provides a stronger osmotic force than the ionic interaction with DMF. The increase in casting solution viscosity facilitates polymer precipitation; meanwhile, the concentration difference in salt solution drives the DMF/water into the polymer-poor phase faster and promote the growth of long well-aligned pores. While the addition of the inorganic nanoparticles such as SiO₂/CeO₂ has negligible effect on the properties of solvent/nonsolvent and thermodynamic process, and results in formation of large macrovoids in the solvent-rich zone and membrane skeleton in polymer-rich zone with the solvent and nonsolvent exchange. Furthermore, certain pore-forming agents could also be used for more hierarchical porous structures. Hierarchical porous frameworks with carbon nanofibers (CNTs) as skeleton and PAN polymer or carbon as the crosslinking shell can be constructed with CNTs as additives (Figure 2F,G). Contrarily, when the mass ratio of incompatible additives (such as CNTs) increases, the lower compatibility between the solvent and nonsolvent significantly reduces the solvent-nonsolvent exchange rate. The nonsolvent-rich phase achieved a certain degree of aggregation before the nucleation of the polymer-rich phase. The dense layer disappeared and uniform sponge-like pores formed in the membranes.

FIGURE 2. Schematic of the regulation of membrane structure by phase separation with different additives

Nonsolvent-induced phase separation

NIPS, known as immersion precipitation, is the most widely used technique in asymmetric membrane fabrication.^36,37 NIPS is suitable for a ternary system, whose main components involve a polymer, good solvent, and nonsolvent, in a homogeneous polymer solution. In NIPS, a casting solution with a suitable concentration is casted on a flat support to form the liquid film. This film is immersed into a coagulation bath which contains a nonsolvent for the polymer solution. A gradient of chemical potential, which is a large driving force, is applied to exchange the solvent and nonsolvent. This is accompanied by mass transfer, phase separation, and polymer solidification. Vapor-induced phase separation (VIPS) and liquid-induced phase separation (LIPS) are also included in NIPS. Jiang et al. systematically investigated the impact of controllable membrane fabrication conditions on the porous structures of membrane, including the polymer concentrations, exposure time in air, solvent compositions, and nonsolvent kinds.³⁸ Polymer solutions of ethyl cellulose-g-poly (2-[dimethylamino] ethyl methacrylate) (EC-g-PDMAEMA) were cast on a precleaned planar glass substrate using a doctor blade. After the so-called “open time” in contact with air (Figure 3A, relative humidity was 30%–40%, 20°C), the as-cast films were immersed into a bath containing deionized water for final formation of the membrane morphology (Figure 3B). The sizes of the membrane pores reduce with increasing connection of tetrahydrofuran in the solvent mixture from 25% to 50%. The solvent compositions affect its evaporation and extend the solution extraction time, thus delaying the phase separation, suitable to produce fine holes. However, with prolonged exposure, the solvent evaporates and the polymer chain further expands at the polymer-air interface, thus eliminating the polymer diffusion and mass transfer rates during NIPS. As a result, the pore structures become small and uniform over the exposed time in air, as shown in Figure 3C–F.

FIGURE 3. SEM images of the cross-sections of EC-g-PDMAEMA and PVDF membranes synthesized under different condition.38,39 (A) Schematic of VIPS, (B) liquid-induced phase separation, (C–F) SEM images for EC-g-PDMAEMA membranes after different “open time”: 10, 20, 40 s, 2 min, respectively, (G–I) cross-section SEM of PVDF membranes prepared at different polymer concentration, 10%, 15% and 20% respectively, (J, K) SEM of PVDF membranes cast at 25°C and 60°C, (L, M) upper and bottom parts of the membrane cast at 60°C, (N, O) PVDF membranes at different exposure time, 15 min and 3 h respectively, (P, Q) PVDF membrane at different precipitation temperature of 25°C and 60°C respectively

Buonomenna et al. prepared polyvinylidene fluoride (PVDF) porous membrane by wet and dry wet phase inversion process.^39,40 The effects of several parameters such as precipitation temperature, composition of the polymer solution (concentration, type of solvent), exposure time before immersion in coagulation bath were studied. From Figure 3G–I, the influence of polymer concentration in the casting solution on morphology is relatively small. The skin layer is progressively tighter and nonporous. Underneath the skin is a region composed of parallel columnar macrovoids. Finally, the lower part of the membrane cross-section shows a cellular morphology. Actually, the dry wet phase inversion process is very convenient for tuning the porous structure of membranes by varying the preparation conditions such as temperature of the casting solution (Figure 3J–M), exposure time before immersion in water (Figure 3N,O), and precipitation temperature in nonsolvent (Figure 3P,Q).

Thermally-induced phase separation (TIPS)

The types of solvents and nonsolvents are significant for the casting solution properties as they critically affect the morphology of the fabricated membranes. The compatibility of the polymer and solvent determines the polymer selection. For crystalline or insoluble polymers, TIPS is a suitable method for the commercial membrane preparation with immense superiority in a wide range selection.^41,42 In TIPS, the solution is heated to a certain temperature above the melting point of polymers, which is named as the upper critical solution temperature, to obtain a homogenous casting solution. Subsequently, the phase separation takes place along with the solution cooling. When the cooling is determined by the liquid–liquid phase demixing, it tends to create the cellular or bi-continuous structure instead of the spherulitic structure, which is dominated by the liquid–solid demixing.^43–46 The extra solvent seems to be the diluent that requires unnecessary extraction to be removed. Benefiting from only two components, TIPS easily operates with less number of defects and randomness. Notably, the formed membranes exhibit a high flexibility, large porosity, and good mechanical properties.⁴⁷ Xu's group have fabricated PAN membranes with controlled pore structures by TIPS, and the sheet-, needle, and cellular-like pores are shown in Figure 4, which also prove that solvent composition, polymer concentration, and cooling rate are significant factors for controlling the membrane morphology.⁴⁸

FIGURE 4. SEM images of PAN membranes prepared with different conditions: (A–D) different PAN concentrations with 30 wt.% glycerol content in water bath at 30°C, (E–H) 12 wt.% PAN with different glycerol contents in water bath at 30°C, (I–M) different cooling baths with various temperatures

Along with the inevitable complexity of solvent extraction and energy consumption at elevated temperatures, it is also a significant challenge to select ideal diluents in TIPS. High boiling point, good thermal stability, and low toxicity are indispensable during heating.^49–51 Besides, the low molecular weight is beneficial to induce interconnected pores with higher porosities. It is crucial to adjust the pore structure parameters for improving the membrane performance, such as the pore size, pore distribution, and porosity. Apart from controlling these conditions, many researches have focused on the optimization of other preparation parameters. Cao et al. applied NIPS to adjust the morphology of porous PVDF membranes using dual-coagulation compositions.⁵⁰ Lee et al. reported a green and sustainable method to prepare porous PVDF membranes that combined NIPS and TIPS, and further investigated the thermodynamic and kinetic effects.⁵¹ In summary, the membranes can be prepared by phase inversion, and the main technologies used for the membrane fabrication in LRBs are TIPS and NIPS. A comparison of these is summarized in Table 1.

TABLE 1 A summary of common phase separations in LRBs

[Table omitted. See PDF]

Electrospinning

Electrospinning is an efficient technique to prepare the membrane with ultrafine nanofibers with varied diameters from micron to millimeter scale. As shown in Figure 5A, the electrostatic technique consists of a grounded collector, syringe containing polymer solution with a stainless steel needle, and high voltage supplier, which is applied to provide the electric field as electrostatic repulsive force. In typical electrospinning, the morphology of polymer solution at the tip of the capillary tube is influenced by the function of the surface tension and electrostatic repulsion.^56–58 As the electric field strengthens, free charges gathered on the surface of the droplet induce distortion. When the electric field achieves a critical value, a Taylor cone will be formed. With further strengthened electric field until the repulsive force is higher than the surface tension, a charged jet is ejected from the fluid to rotatory collector. Accompanied with solvent evaporation, the polymer nanofibers are collected on the collector. The obtained nanofiber membranes possess the advantages of interconnection, large surface area, and high porosity. To prepare the core-shelled or/and even multi-core structure, coaxial and multi-spinneret are performed using different collectors.^59–62

FIGURE 5. (A) Schematic representation of electrospinning equipment,52 (B–E) SEM image of gelatin nanofibers prepared with various polymer concentrations,53 (F–I) effect of voltage on morphology and fiber diameter distribution.54 (J–M) SEM images of E-spun PLLA fibers prepared at different temperatures55

Furthermore, the unique features of fiber diameter and its morphology are significantly affected by the following controllable parameters: (1) the system parameters such as molecular weight and polymer solution properties (solubility, viscosity, conductivity, and dielectric constant). As shown in Figure 5B–E, an increasing concentration, causing an increase in viscosity, necessarily results in the formation of bigger diameter fibers.⁵³ Large polymer molecular weight and low conductivity of the solution also tend to produce bigger fibers. (2) Processing parameters, such as the electric potential, flow rate, and distance between the capillary and collection screen. It has been proved that, there is a slight increment in the average fiber diameters when the applied electric field is increased, as shown in Figure 5F–I.⁵⁴ The flow rate and distance are related to the solvent evaporation time. An extremely high flow rate and a short distance between the tip and collector cause bead formation because of the insufficient time for evaporation, while the low flow rate and long distance fail to form continuous fibers. (3) Ambient parameters, such as temperature and humidity. When air humidity is low, the evaporation rate is higher than the removal of solvent, which clogs the polymer in the spinneret tip, while the high humidity entraps the moisture in the fiber membrane, tending to form the porous structure. A high temperature can generate thin fibers (Figure 5J–M).⁶³

Zhu et al. investigated the influence of the spinning solution viscosity on carbon nanofiber diameters, as shown in Figure 6A.⁶⁴ When the viscosity was enhanced by increasing the coal-based graphene quantum dot (CGQD) content in the spinning solution, the oxygen-bearing functional groups provided a strong interaction with PAN molecules, resulting in the increasing diameter of carbon nanofibers and ensuring the formation of dense, strong, and flexible carbon skeleton. A two-layer electrospun membrane was uniquely designed,⁶⁵ as shown in Figure 6B, which was composed of a thin top CNT/polymer nanofiber layer and thick bottom porous neat polymer layer. The incorporation of micro/nano beads with CNT increased the surface roughness and provided the desirable characteristics. To construct highly functional and high-performance nanofibers for the challenges in application, inorganic additive and salt doping have been studied to analyze their influence on the morphology and performance of the membrane. The addition of salts leads to a higher charge density and reduces the final fiber diameter. Furthermore, the peculiar bubble-nanorod structure Fe₂O₃C composites, SiNP@C core-shell fibers, are also created by electrospinning, as shown in Figure 6C–G. Han et al. developed a core−shell structure by incorporating poly(vinyl alcohol) (PVA) and poly(acrylic acid) (PAA) based on the electrospinning and in situ polymerization coating (Figure 6H).⁶⁸ The cross-linked nanofibrous membrane exhibits a large porosity, enhanced specific surface area, superior tensile strength, and superior mechanical flexibility for the application of flexible materials.

FIGURE 6. (A) Schematic of carbon nanofiber preparation,64 (B) cross-sectional SEM image of a two-layer electropun membrane,65 (c) schematic of bubble-nanorod-structured Fe2O3C nanofibers, (D) TEM images of Fe2O3C nanofibers,66 (E) SEM of SiNP@C nanofibers, (F, G) electrospinning preparation of core-shell fibers with Si nanoparticles,67 (H) Schematic of core−shell cross-linked nanofibrous membrane68

3D printing

3D printing is an emerging advanced manufacturing technique, also called as additive manufacturing. Since the concept of 3D printing was introduced in the 1980s, it has drawn significant attention and research in recent years and appears promising in the fabrication of industry applications. Compared to the traditional manufacturing methods, 3D printing has particular advantages, including low cost, flexible manufacturing, wide selection of materials, geometry designs, and rapid prototyping.

Constructing a virtual model is the premise for 3D printing technology, a three-dimensional object with geometric periodicity is created using the computer-aided design (CAD) software. Then the surface information of the model is converted to the STL (StereoLithography) file format, which is the storage format for the model's two-dimensional coordinate. Finally, the printer successively deposits materials of such 2D layers and solidifies them to add layer upon layer until the object is completely created. As a more accessible version, 3D printing technology is achieved by the transformation between 3D object models and virtual geometries. Based on the computer-aided data design, materials are accumulated into a physical structure of the manufacturing procedures. 3D printing is a series of techniques, and not a single technique. According to each fabrication process and modeling principle, 3D printing techniques can be divided into the following categories, as shown in Figure 7A, defined by ASTM International in 2015.⁶⁹ As a novel manufacturing technique, 3D printing has a wide range of applications, such as bioscience,⁷³ medical,⁷⁴ manufacturing and engineering, as well as membrane systems.^75,76 A 3D printed polysulfone (PSU) membrane prepared by selective laser sintering was applied in the oil/water separation,⁷⁷ Wessling's group directly printed polydimethylsiloxane (PDMS) for a gas–liquid contact by 3D-printing in a rapid prototyping device.⁷⁸ Freestanding carbon nitride-based hybrid aerogel membranes were prepared based on a 3D printing technique (Figure 7B), which assembled 2D materials into large-scale 3D systems.⁷⁰

FIGURE 7. (A) Schematic illustration of 3D printing techniques and their respective definitions,69 (B) the schematic illustration of the fabrication of carbon nitride-based hybrid aerogel membranes,70 (C) schematic representation of 3D printing techniques, (D) schematic illustration of printing LFP and LTO with micro-lattice architectures,71 (E) schematic of 3D-printed SnO2 QDs/G electrodes72

Compared to the complicated and expensive traditional manufacturing techniques, 3D printing can easily fabricate the objects with complex geometry by a well-controlled layer-by-layer deposition. Furthermore, the complexity of the objects with great versatility and reduced prices promote the application of 3D printing in the energy storage.^79–82 Carbon-based materials (graphene-based, carbon nanotubes, etc.) with high surface areas and well-designed periodic microlattices were printed by 3D printing.^83–85 Furthermore, significant achievements have been made by fabricating LTO/LFP/LMO/LCO-based materials with interdigitated microbattery architectures,^82,86,87 such as LFP and LTO electrodes with micro-lattice architectures using 3D printing techniques, as shown in Figure 7D.⁷¹ In addition, many researches revealed that 3D printing techniques are facile and convenient for material manufacturing and designing as components of batteries, including electrodes, electrolytes, and separators.^88–90 SnO₂ quantum dots, graphene oxide, and deionized water were mixed into the printing ink to fabricate well-designed periodic micro-lattice architectures as electrodes with a large area and space.⁷² Considering further applications and large-scale fabrication, important properties (shown in Figure 7C) must be involved for the suitability of the 3D printing techniques. To further improve the performances of 3D printed applications, the construction refinement needs a higher resolution, accuracy, and controlled build size. Also, the compatibility and scalability of printable materials and mechanical properties of printed objects are important, which ensure that the products remain stable and available on target. When these conditions are satisfied, advantages like quicker print speed and low cost can be achieved in the development of 3D printing techniques.

Other methods

Other methods have also been used for the preparation of porous membranes, such as freeze drying,⁹¹ breath figure method,⁹² stretching method, and interfacial polymerization,⁹³ which exhibit the desirable properties in membranes and flexibility for further modifications. Considering the manufacturing procedure and performance requirements for diverse applications, the fluids in composite membrane have gained an increased attention in self-supporting membranes. The crucial advantages of the self-supporting membranes include a high mechanical stability and flexibility, accurately controlled thickness with nanochannels, as well as modified properties with the introduction of doped composition.^94,95 Our group prepared PAN doping Fe₃O₄ membrane by freeze drying, as shown in Figure 8A. Oriented and aligned pores with connected network were generated under the action of directional temperature gradient. After carbonization, Fe₂C hybrid nitrogen-doped carbon (Fe₂C/NC) membrane possessed highly connected and regular conductive networks and provided fast diffusion channels in the energy storage applications. g-C₃N₄ nanosheet membranes with self-supporting spacers were synthesized. The nanochannels generated by the self-supporting spacers in the g-C₃N₄ membrane provided stable and rigid structures. The artificial nanopores and spacers between the nanosheets enabled an efficient molecular transport.⁹⁶ A self-supporting flexible and transparent graphene film with wrinkled-graphene-assembled open-hollow polyhedron building units was developed using chemical vapor deposition. Such a self-supporting structure facilitated the transfer of the film onto any desired substrate without damage and widened the practical applications. The face-to-face mode structure also enhanced full connectivity, along with the mechanical properties of membranes.⁹⁷ The composite membrane can also be composed of inorganic particles. Doped ZnO seed-nanoparticle polymeric membranes were prepared by phase separation, which enabled a pore-specific in situ growth of small (<5 μm in diameter) ZIF-8 islands, as shown in Figure 8G,H.⁹⁸ Wang et al. suggested a simple method for constructing hydrogel “bridges” for freestanding CaCO₃ films,⁹⁹ Figure 8J shows CaCO₃ microparticles connected to a silk-like “adhesive,” acting as a spider web. Hydrogel bridges dragged the rigid CaCO₃ particles together, enhanced the flexibility of multifunctional membranes, and provided abundant freedom space among the inorganic particles. The detailed fabrication process is shown in Figure 8K.

FIGURE 8. (A) Scheme of Fe2C/N-C membrane prepared by freezing-drying, (B) diagram of self-supporting g-C3N4 nanosheet membranes via artificial nanopores,96 (C) schemes of synthesized self-supporting graphene membranes, (D) sectional view of two superimposed graphene membranes, (E) SEM image of graphene membranes with open-hollowed building units, (F) bottom view of graphene membranes,97 (G) formation of ZnO seed-nanoparticle membranes, (H) SEM image of supported ZIF-8-PES composite membrane cross-section,98 (I) SEM image of CaCO3 film cross-section, (G) SEM image of Ca-Alg gel acting as organic adhesive, (K) proposed mechanism for the formation of CaCO3 film

APPLICATIONS IN LRBS Separators

The basic functionality of a separator is avoiding the actual anode-to-cathode contact without any direct cell reactions and providing channels for Li⁺ flow. Although it is an inactive component during the charge or discharge of batteries, the separator is an essential medium for Li⁺ transport. The structure and property significantly influences the performance and durability of a battery. In particular, the formation of a solid electrolyte interface (SEI) layer and growth of Li dendrites give rise to short circuits and electrode pulverization, resulting in serious safety issues.

Ideal separator characteristics

Separators must be stable to endure the extremely strong oxidizing and reducing conditions for guaranteed safety throughout the charge and discharge processes. Simultaneously, a low internal resistance and high ionic conductivity are desired to achieve a high capacity and long cycling performance. Figure 9 summarizes the general requirements for the applications in LRB separators, and the following section provides a detailed discussion.

FIGURE 9. Schematic of the ideal characteristics of high-performance separators

Safety is an important factor in the practical application of batteries. Mechanical strength and thermal stability are crucial to sustain safe performance in long operation periods. The separators must possess a high mechanical strength, including the machine and transverse directions to bear the tension of winding operation and electrode deformation in the battery assembly and usage, thus avoiding internal short circuits. When the battery is charged, especially overcharged, hazards may arise due to the thermal shrinkage of separators. Therefore, an ideal separator must possess thermal stability at relatively high temperatures. In general, the low polarity of separator materials causes poor affinity and inferior wettability of the electrolytes, while the appreciable permeability of the separator can provide the high wettability and electrolyte uptake, which results in efficient ion transport and decreases the resistance. The structure of membrane is an important characteristic, affecting the safety and electrochemical performances. The separators should possess the appropriate pore size and porosity. The low porosity increases resistance, which is unfavorable for the electrolyte wettability. However, an extremely high porosity fails to block the active materials and maintain the mechanical strength, adversely impacting the safety. Usually, the appropriate pore size for separators is less than 1 mm. In addition, electronic insulator, suitable thickness, and chemical stability are also desirable.

Polyimides (PI) are one kind of high-performance polymers on account of its excellent chemical resistance, good thermal stability, and superior mechanical properties. Sun et al. have fabricated ZrO₂-armored hybrid separators (ZrO₂@PI) by in situ growing ZrO₂ nanolayer on polyimide (PI) nanofibers with the aim of improving the fire resistance, thermomechanical properties, wettability, and safety of LIBs (Figure 10A). ZrO₂ shell layers are formed on the surface of nanofibers to markedly promote the wettability, structural stability, and strength of PI nanofiber membranes.¹⁰⁰ The ZrO₂@PI separators can maintain superior thermal-dimensional stability up to 360°C, and this is superior to that of the initial PI (320°C) and polyolefin separators (140°C). In addition, a robust flame retardant fluorinated polyimide (F/PI) nanofiber separator has been designed for high-temperature Li-S batteries (Figure 10B,C).¹⁰¹ Cui et al. have fabricated a novel electrospun core-shell microfiber separator with thermal-triggered flame-retardant properties for lithium-ion batteries (Figure 10D).¹⁰² The microfibers exhibit a core-shell structure, where the triphenyl phosphate (TPP), a popular organophosphorus-based flame retardant, is the core and poly(vinylidene fluoride-hexafluoropropylene) (PVDF-HFP) is the shell. The encapsulation of TPP inside the PVDF-HFP protective polymer shell (TPP@PVDF-HFP) has prevented direct exposure of the flame retardant to the electrolyte. The PVDF-HFP exhibits a relatively low melting point (~160°C), such that it can be melted before or at the early stage of combustion. By using the flame-retardant properties of MoO₃ and the excellent Li⁺ conductivity of LLZTO, He et al. developed a robust bi-layer separator by incorporating MoO₃ and Al-doped Li_6.75La₃Zr_1.75Ta_0.25O₁₂ (LLZTO).¹⁰³ The bi-layer separator is highly flame-resistive, and manages to endure intense fire (Figure 10E), which owns a large ductility (227%), and a low thermal shrinkage (5%) after annealing at 160°C for 4 h, and exhibits simultaneously excellent mechanical, wetting, thermal shrinkage, ion conductivity, thermal distribution and electrochemical performances. Moreover, a nonwoven ZrO₂ ceramic membrane with a robust nanofiber microstructure via polymeric electrospinning followed by a high-temperature organic burn-off has been reported (Figure 10F,G).¹⁰⁴

FIGURE 10. (A) Fabrication process of ZrO2@PI membrane via in situ reinforcing technique, (B, C) the F/PI nanofiber separator, (D) fabrication of the microfibers by electrospinning and SEM image of the membrane, scale bar is 5 mm, (E) the cross-sectional SEM images of bi-layer separator and the schematic illustration for the bi-layer structure, (F) HRTEM image of the ZrO2 membrane, (G) fire-resistant tests of the ZrO2 separator

In addition to the safety function of a battery separator, the interplay between the mechanical and electrochemical properties is also key selection criteria. Freunberger et al. have investigated the mechanical properties of the conventional PP membrane using tensile and puncture penetration tests at abuse relevant conditions. In order to consider the often strong anisotropic mechanical behavior caused by the production process, three specific rectangular specimens are created. Specimens are cut in the main direction in machine direction (MD), transverse direction (TD) and at an angle of 45°to the MD (Figure 11A,B).¹⁰⁵ Equally the tensile behavior was normalized for the TD and 45°directions. The normalized stress levels of the selected discrete strain values (2.0%, 4.0%, 6.0% and 8.0%) for TD and 45°and the arithmetic mean value over all selected percentage stress values for each separator are shown in Figure 11C. Compression stress on thickness direction of separators exists in assembling and operation of all lithium-ion batteries. Compression was found to be key factor influencing the microporous structures, electrolyte uptake and electrochemical properties of separators (Figure 11D–F).¹⁰⁶ Xiang et al. reported a Sb₂O₃ modified PVDF-CTFE electrospun fibrous membrane (SPCF) as a safe lithium-ion battery separator.¹⁰⁷ Besides, the Sb₂O₃ modified PVDF-CTFE electrospun fibrous membrane exhibited synergistic flame retardancy as well as outstanding thermal stability without shrinkage at 160°C for 2 h. The maximum stress and strain of the PCF membrane are about 4.6 MPa and 87.5%, while those of the SPCF membrane are about 13.5 MPa and 141.1%. These improved mechanical properties are attributed to the Sb₂O₃ ceramic nanoparticles inside the nanofibers (Figure 11G,H). Gennady et al. have investigates the effects of electrolyte solvents on the mechanical properties of a polypropylene battery separator through experimental measurements of thickness and elastic modulus of separator samples immersed in different solvent environments (Figure 11I).¹⁰⁸

FIGURE 11. (A, B) Orientation of the specimen in relation to the machine direction for tensile test and specimen dimensions for electrochemical test, (C) normalized arithmetic stress mean values for tensile tests in MD, TD and 45°over all directions (MD, TD and 45°), (D, E) nominal stress–strain curves and stress–strain recovery curves of the three separators (through-thickness compression), (F) increase of electrolyte level at wetting time of 90 s of the separators before and after compression, (G) stress–strain curves of the PCF membrane and SPCF composite membrane in comparison with commercial PE membrane, (H) SEM images of the SPCF composite membrane, (I) Young's modulus for separator samples immersed in different solvents as a function of separator thickness

Membrane separators and modification

At present, the commercial microporous polyolefin separators over the last few decades, such as polyethylene (PE), polypropylene (PP), and PE/PP/PE separators, provide the advantages of good mechanical strengths, chemical stability, and low cost. However, the weak wettability resulting from the low porosity of polyolefin and inferior thermal stability constrained the development of LRBs owing to the safety hazards and poor capability. The effective method for improving the performance of polyolefin separators is surface modification, which can be classified into surface grafting and surface coating. The addition of electrolyte-philic monomers with polar functional groups and ceramic materials can increase the polarity of membranes, facilitating the liquid electrolyte wettability and suppressing the growth of lithium dendrites. The downside of the design is the thick coating layers that inevitably block the Li⁺ transfer and increase the battery resistance with depressed energy density.

Different preparation methods result in membranes with various unique structures. The stretching step can form porous structures. Plentiful researches have proved that slit-like pores (Figure 12A) prepared by the dry process are suitable for high power density, while the interconnected and elliptical pores (Figure 12B) prepared by the wet process tend to go for a long cycle life.^109,110 Liu et al. produced a cross-linked polybenzoxazine electrospun fiber mat (CR-PBz-FbM) porous separator by electrospinning. The membranes possess stable mechanical and thermal properties.^111–114 Cho et al. obtained PAN membranes with various pore structures and controlled thicknesses by electrospinning.^115–118 To improve the original function of a PAN membrane, other than low density and good flexibility, Kim et al. designed the difunctional polymeric separators by a dual-nozzle electrospinning. These membranes composed of agarose, PAA, and PAN, act as separators and play a role in suppressing the Mn dissolution of LiMn₂O₄ cathodes, as shown in Figure 12C.¹⁰⁸ Hu's group reported a thermal management separator combining boron nitride (BN) nanosheets with poly vinylidene fluoride-hexafluoropropene (PVDF-HFP) for LRBs, as shown in Figure 12D,E.Figure 12F–J shows that the membranes consisted of continuous porous structures with a thickness of approximately 30 μm, which is suitable for a fast Li⁺ transport. The improved electrochemical performance with high coulombic efficiency in Figure 12K can be ascribed to a uniform thermal distribution interface, enabling the homogeneous Li nucleation and suppressing the Li dendrite growth.

FIGURE 12. (A) SEM images of microporous membrane separators prepared by dry process (B) SEM images of microporous membrane separators prepared by wet process, (C) illustration depicting the positive effects of the (agarose/PAA)/PAN coaxial electrospun NF separator, (D) schematic illustration of BN-PVDF-HFP separators by 3D printing, (E) digital photographs of BN-PVDF-HFP separators, (F–J) SEM images of BN-PVDF-HFP separators, (K) coulombic efficiencies of the Li/Cu cells of BN-PVDF-HFP separators at 1 mA cm−2, (3D printed separator for thermal management of high-performance Li metal anodes), (L) SEM images of PVDF/PAN separator, (M) photographs of the separators before and after heat treatment at 160°C, (N) rate performance of lithium-ion batteries (PVDF/PAN blend separators via thermally-induced phase separation

Semi-crystalline PVDF polymer is an attractive candidate for a separator owing to its higher polarity. Phase inversion technology is an effective method to fabricate porous PVDF membranes as separators. TIPS is the most common method because the PVDF membranes prepared by TIPS exhibit a higher ionic conductivity and stability. It is also a universal procedure for effectively introducing the electrolyte-philic materials to enhance the wettability of the separators, either polymers or inorganic particles (inorganic salts).^109–115 Wu et al. fabricated PVDF/PAN blend separators by TIPS and the SEM is shown in Figure 12L. Separators with rod-like pores showed less shrinks after a heat treatment at 160°C for 1 h, as shown in Figure 12M, verifying the superior thermal stability with the introduction of PAN. Moreover, the addition of PAN could affect the pore structures of PVDF/PAN blend membranes related to the ionic conductivity. As a result, the batteries using blended separators exhibited a better rate performance, as shown in Figure 12N.

Interlayer in Li-S batteries

Lithium-sulfur batteries have received considerable attention because of their high theoretical specific capacity. However, quick capacity decay and severe shuttle effect are extremely necessary challenges to be addressed. A separator for Li-S batteries should be able to suppress the polysulfide migration and protect them from dissolving into the electrolyte. Manthiram first proposed the concept of interlayers in 2012 by inserting a carbon-based film between the cathode and separator,¹¹⁶ which immensely inhibited the dissolution of polysulfide and enhanced the cycling stability. Polar barriers are the desired materials with the strong chemical adsorption of polysulfides, such as metal oxides, metal sulfide, and heteroatom doping functional carbon.^117–120 PAN- and PI-based CNFs were prepared by different techniques, as the interlayers were reported.^121,122

Li et al. extended the phase inversion into the applications of interlayers (Figure 13A).¹¹⁹ The cross-linked mesoporous and patterned macroporous Fe₃CC/CNT membranes were fabricated. The membranes exhibit flexibility (Figure 13B,C) and high adsorption of polysulfide (Figure 13E–G). The highly efficient polysulfide trapping and fast lithium-ion diffusion could be realized by the unique pore structure with better performance. SEM of Fe₃CC/CNT interlayer after cycling was also studied to reveal the adsorption of hierarchical porous networks. The patterned channels retained favorable structures even after cycling with a high sulfur loading of 7.0 mg cm⁻², which prevented the transport blocking of Li⁺ ions (Figure 13H–J). With the insertion of Fe₃CC/CNT interlayer, the cells enhanced the electrochemical kinetics and exhibited a lower resistance, as shown in Figure 13K. The areal capacity reached 3.81 mAh cm⁻² after 100 cycles when the sulfur loading was up to 7.1 mg cm⁻², providing an effective manufacturing strategy for fabricating the interlayers.

FIGURE 13. (A) Illustration of battery structure with interlayers and patterned channels for highly efficient polysulfide trapping and facilitated Li+ diffusion, (B, C) photos of Fe3CC/CNT membrane, (D) hierarchical porous networks of membranes, (E) Li2S6 solution adsorption with different membranes, (F) U-type permeation device with Fe3CC/CNT membrane as an interlayer, (G) UV–vis spectrum, (H–J) SEM images of Fe3CC/CNT interlayer after 100 cycles at 0.2 C with high S loading, (K) EIS spectra of cells with and without Fe3CC/CNT interlayer, (L) areal capacities of the cells with Fe3CC/CNT interlayer and different S loadings at 0.2 C119

New flexible carbon membrane interlayers consisting of carbon nanotubes (CNTs) as interlayers to trap the lithium polysulfides were prepared, as shown in Figure 14A,D. The macropore network formed by the entangled matrix of CNTs accommodated the liquid electrolyte and improved ion conductivity.^120,122 The cell with a PVDF-HFP + CNF interlayer exhibited higher discharge/charge capacities and a smaller degradation rate of 0.092% per cycle with the loading of 1.7 mg cm⁻² (Figure 14K). A three-dimensional (3D) reduced graphene oxide/activated carbon (RGO/AC) interlayer was developed using a simple hydrothermal synthesis and convenient mechanical pressing; the structure of the membranes is shown in Figure 14E–J.¹²¹ An MOF membrane was also applied as an interlayer to suppress the polysulfide shuttling in Li-S batteries. The crystal structure of Cu₂(CuTCPP) is shown in Figure 14B. Vacuum-assisted filtration was used for the Cu₂(CuTCPP) membrane (Figure 14C) with high orientation, controlled thickness, large lateral size, and good flexibility.¹²³

FIGURE 14. (A) Principle of reduced shuttle phenomenon with GO/CNT membrane as an interlayer,120 (B) simulated crystal structure of Cu2(CuTCPP), (C) SEM image of Cu2(CuTCPP) membrane,123 (D) pristine carbon nanofiber interlayer after discharge processes, (E–H) SEM images of RGO/AC interlayers, (I) TEM images of RGO/AC interlayers, (J) schematic of RGO/AC interlayers,121 (K) long cycle performance of batteries with a PVDF-HFP + CNF interlayer122

Porous membrane based electrode designs

Significant attention is being given to portable applications, such as electronic devices, which increases the demand for energy density and specific capacity.^109,110 It is necessary to reduce the relative weight and volume of inactive components.¹²⁴ In major configuration LRB batteries, the electrodes play a significant role in improving the electrochemical active material ratio in the battery configuration. The structural engineering of thick electrodes, related to the membrane electrodes, alleviates the issues and makes the battery more efficient.

Design principles for fast ion transport

Conventional electrodes consist of conductive additives (carbon black), polymer binders, and active materials. If we focus on thick electrode by thickening the coating nano powders in layers (Figure 15A), followed by drying the slurry, the cracks and delamination of active materials are inevitable. Moreover, the high electrical/thermal resistance and tortuous electrons/ions diffusion paths in the electrode restrict the electrochemical performance. In addition, poor adhesion between the thick electrodes and current collectors is also a problem for thick electrodes.^111–114 Tortuosity (τ), is the only valid parameter to reflect the complexity of the transport in a porous network of electrodes, which is assumed to affect the electrode performance. As described in Equation 1,[Image Omitted. See PDF]where ε is the porosity, D is the intrinsic diffusion coefficient, D_eff is an effective diffusion of lithium in the electrolyte, represents the effective ionic conductivity.^115–118 Consequently, it is believed that low-tortuosity pore structure enhances the ion transfer as a vital design principle for thick electrode architectures.

FIGURE 15. (A) Schematics of ion transfer pathways in diverse electrodes, (B) Schematic of electrode by assembling stack, (C) SEM images of a cross-section of MWCNTs on 3D Cu mesh, (D) SEM image of VACNT-based thick electrodes on Cu foil current collector, (E) formation mechanism, preparation and ion/electron transport illustration of membrane electrodes, (F) optical images of membrane electrodes, (G) optical images of traditional electrodes with increasing sulfur mass

Low-tortuosity electrodes can be achieved by subtractive and additive design approaches. As the name suggests, subtractive design involves the formation of the aligned pore structure followed by subtracting certain components from the electrode precursor during the phase inversion or electrospinning process.^125–127 The additive design approach to fabricate the electrodes by adding layer upon layer of the slurry, such as in 3D printing.⁸² Based on the principle for electrode design, a conclusion can be drawn that a low-tortuosity structure with a straight and aligned pore design can contribute to enhanced ion transfer kinetics.

Modified thick electrodes

Kang et al. developed novel 3D MWCNTs on Cu current collectors as an anode of LRB, as shown in Figure 15B. The enhanced specific capacity can be attributed to the higher average solid loading of MWCNTs brought by the thick electrodes.¹²⁸ Kara et al. reported an alternative scalable method to produce ultrathick aligned carbon nanotube electrodes with high conductivity and stability in LRB battery. The interconnected structure (Figure 15C) and binder-less electrodes boost the electrical and thermal conductivities, giving rise to improvements in high-performance batteries.¹²⁴ Zhang's group successfully prepared high-performance porous electrodes via phase inversion. The electrode slurry was coated to aluminum (Al) foil and the phase inversion took place when the foil was immersed into a water coagulation bath driven by the broken thermodynamic equilibrium and existence of the chemical potential gradient. Furthermore, the polymer-rich phase produced the skeleton of porous electrodes, and polymer-poor phase formed the pore structure. The connected pores provided convenient channels for ion transport and the illustration is shown in Figure 15D,E. Additionally, the single component of electrode materials without binder offered a stronger Van der Waals force and enhanced the adhesive strength. Figure 15F,G show that the electrodes were strongly cohesive on the Al foil with more sulfur loading.

In fact, metal foils as the current collector only work on collecting the current generated by the active materials to offer a larger current output, not belonging to the indispensable active materials, and have its downside for increasing internal resistance and overall weight of the batteries. Previously, He et al., proposed a universal method to prepare the highly scalable yet flexible asymmetric porous structured membrane by phase inversion. Various asymmetric pores can mostly satisfy the demands of battery performance. More importantly, the asymmetric porous structured membrane with a dense layer can act as an active material and current collector, avoiding the use of separate current collectors, even conductive agents and binders in lithium-ion battery, which is beneficial for superior electrochemical performances in terms of high reversible capacity.

The exchange rate of solvent and nonsolvent is crucial for fabricating the various pore structures of membranes. Dense layer or sponge-like macroporous triple-layer-structured membrane and dense layer/finger-like pore/honeycomb-like pore asymmetric porous membrane have been successfully synthesized with the addition of polymeric additives (poly[ethylene glycol], PEG), as shown in Figure 16A–J. The polymeric additives increased the solution viscosity, affecting the phase separation kinetics to obtain the diverse pore structure.¹²⁹ Besides, inorganic doping could be realized by adding the metal salts and nanoparticles into the casting solution. Different colors of the membranes are shown in Figure 16K–N.^129–133 Subsequently, the membranes were cut into the circle as integral electrodes, as shown in Figure 16O, and served as the host with large S loading (Figure 16Q) showing promising candidates for future commercial applications in Li-S batteries with more than 30 lighted blue light-emitting diodes (Figure 16P). Figure 16R,S show electrochemical performances of these porous membranes as the electrode material. Fe₃C/C membrane electrodes exhibited the specific capacity of 323.3 mA h g⁻¹ after 250 cycles in LRB battery. When the SiO₂/C membranes were used in Li-S battery, the average specific capacities of 1050, 935, 855, 767 and 649 mA h g⁻¹ were measured at 0.1, 0.2, 0.5, 1 and 2 C (1 C = 1675 mA h g⁻¹) respectively, which are higher than that for traditional Al cathodes and C membranes without SiO₂ doping. Moreover, the reversible capacity of 927 mA h g⁻¹ was obtained when the current density returned to the initial value of 0.1 C. To further improve the areal sulfur loading, the group reported a facile layer-by-layer stacking of Fe₃C/C membrane from one to three layers for high-areal-capacity Li-S batteries, and the sulfur loading reached 4.3 mg cm⁻² in the unique multilayer cathodes with aligned macropores.

FIGURE 16. (A–C) SEM images of cross-sections, (D, E) SEM images of the surfaces of Fe3C/C dense layer/finger-like pore/honeycomb-like pore asymmetric porous membrane; (F–I) SEM images of the cross-sections of Fe3C/C dense layer/sponge-like pore/macropore triple-layer-structured membrane, (J) SEM images of membranes with high S loading; (K) Fe3+-doped brown PAN membranes; (L) Fe2O3-doped red PAN membranes; (M) SiO2-doped white PAN membranes; (N) Carbon nanotube-doped black PAN membranes; (O) membranes cut into integral electrodes; (P) LED lamps lit by two coin cells; (Q) Schematic of membranes as S host in Li-S batteries; (R) cycling performance of Fe3C/C membrane electrodes in LRB battery; (S) rate capability of SiO2/C membrane electrodes in Li-S battery with the sulfur loading of 1.5 mg cm−2

Fast ion transport channels in integrated electrodes

As mentioned earlier, the design of low-tortuosity electrodes with aligned pore is the key to develop high active mass loading, and is not based on the premise of increasing electrons/ions diffusion resistance. Herein, Li et al., devoted to design the asymmetric porous structured membranes as an ideal electrode material, which appears to be a prospect of development. The reason for the excellent electrochemical performances of the membrane electrode requires deep attention and reflection. First, the thick membrane electrodes improve the active materials ratio for a high energy density, and pore structure blocks the re-aggregation of particles. Second, the dense layer of membranes can replace Al foil and serve as current collector with high conductivity and mechanical stability. Third, the membrane electrodes with arrayed and low-tortuous pore structure can provide fast diffusion paths and enhance the ion transport. As shown in Figure 17, the traditional C-S slurry with random, close packing of electrode components, usually block the ion transport channels and cause the shorter board of ion transfer, while membrane with long well-aligned pores could considerably shorten the diffusion path of Li⁺ ion.

FIGURE 17. Principles for fast mass transport as integrated electrodes

For porous carbon membranes, the addition of polar bond derived from doping inorganic/organic materials and heteroatom exhibited the favorable reactivity for strong adsorption and significantly reduced the energy barrier and reaction kinetics for polysulfide conversion. Benefiting from the efficient catalytic/adsorption performance, the hierarchically porous networks carbon membranes trapped soluble polysulfide intermediates and induced the opposite diffusion as reservoirs, preventing the active material dissolution into the electrolyte and facilitating the ion/electrolyte transport for fast reaction kinetics based on the trapping-diffusion-conversion mechanism (Figure 17A). The membrane cathodes with aligned channels and hierarchically porous networks significantly promoted Li⁺ and electron transportation based on the design principles of low tortuosity. From single membrane cathode to layer-by-layer stacked structure to circulation groove with the two membranes, such unique membrane electrodes with aligned macropores largely improved the ion/electron transport with good conductivity and simultaneously buffered the volume expansion and retained the whole structure intact (Figure 17B,C). The classical molecular dynamics simulation in terms of the diffusion process of Li⁺ were modeled and calculated, the free energy curve revealed an initial increase and then decrease when the Li⁺ crossed the electrolyte reservoir, confirming the S/C slurry coating layer on Al foil material had an adverse impact on the Li⁺ transfer (Figure 17C). Also, the universal method for preparing membranes with doping additives results in a higher conductivity and strong adsorption with polysulfides, even catalytic performance for the reaction in battery. Our work for scalable membrane fabrication and structural design obtained an improvement in the specific capacity and power performance with the advantages of mass production, which provided promising strategy for practical applications of high-energy Li-S batteries. There are also emerging researches on designing 3D porous current collectors and freestanding flexible electrodes with low cost and abundant plant-derived biopolymer,^134–136 which is renewable, environmentally friendly, and useful for large-scale production. Further methodology is expected to design the flexible membrane electrode architectures in the application of high-energy storage devices.

Porous membrane based electrolyte designs

Research on LRBs is essential to improve the cell safety performance, especially overcharging and short-circuiting that frequently causes a fire hazard or an explosion. It is a promising choice to replace the liquid electrolyte with more stable and safer solid electrolytes, which can provide higher thermal and chemical stability, as well as stronger mechanical strength, avoiding liquid leakage simultaneously. Meanwhile, the stable ion transport and uniform electrodeposition contribute to realizing the high-energy LRBs. There are rapidly growing technological challenges that focus on the ionic conductivity issues, ion mobility, interfacial property, mechanical strength, and preparation. The microporous membranes can also be used as an electrolyte to ensure Li⁺ transport. The common electrolyte membranes are divided into the following categories: solid ceramic electrolytes, solid polymer electrolytes, and gel polymer electrolytes.^137–141 Figure 18A,B shows the pivotal issue for solid-state electrolytes, the highly porous solid-state electrolytes provide the space for active material loading and electrode volume change, which ensure a competent electrode mechanical strength and facilitate Li⁺ transport. The ingenious designs possess a dense separator layer, suppressing the penetration of Li dendrites and efficiently reducing the cell resistance.

FIGURE 18. (A) Schematic of 3D garnet-type SSE framework with integrated separator layer and electrode host layer, (B) schematics of solid-state batteries based on 3D SSE frameworks,142 (C) schematic of the synthesis of LLTO framework composite electrolytes, (D) surface morphologies of LLTO frameworks,143 (E) fabrication of cathode-supported solid electrolyte membrane framework, (F) SEM image of the cross-section of cathode-supported PPAL solid electrolyte membrane, (G) discharge capacities, (H) rate performance, (I) electrochemical impedance plots of conventional and cathode-supported solid electrolyte batteries144

However, the high crystallinity of the commonly used PVDF may hinder the migration process of Li⁺, resulting in inferior ionic conductivity and stability. Its copolymer poly(vinylidenefluoride-co-hexafluoropropylene) (P[VdFHFP]), poly(vinylidene fluoride)-graft-poly(tert-butyl acrylate) (PVDF-gtBA), and doping ceramic fillers with porous structures prepared by electrospinning lead to a high ionic conductivity and electrochemical stability.^145–149 Miao et al. prepared the (PVDF-HFP)-based PSPE incorporating TiO₂ by the phase inversion method, which exhibited the high ionic conductivity and electrolyte uptake with the addition of TiO₂, as well as improved the interfacial capability, suggesting the improved rate capability. Polymethyl methacrylate (PMMA), employed as polymer blends, also improved the interfacial compatibility between the electrolytes and electrodes, such as PAN/PMMA based gel polymer, showing the superiority in reducing transfer and diffusion resistance, decreasing the polarization of batteries.^150–152 PAN had a high mechanical strength and affinity to electrolyte owing to the strong interaction of CN. This seemed as a good solid-state electrolyte in LRBs, which increasingly attempted for the development of flexible batteries. Table 2 shows the common polymer and its respective characteristics. Yu's group developed numerous novel synthetic methods to prepare conductive polymer gels (CPGs) with cross-linked network and 3D hierarchically porous nanostructures (Figure 18C,D), exhibiting considerable performances, such as excellent electrical conductivity, large surface area, tunable structures, and hierarchical porosity for fast transport channel.^{143,153–156} More scientific and technological progress is expected in this promising field.

TABLE 2 Common polymers and their features as electrolytes in LRBs

Material	Feature	Advantage	Disadvantage
PVDF	F atoms	High mechanical strength Good affinity to electrolyte	Form LiF with lithium–ions
PAN	CN groups	High ionic conductivity Good electrochemical stability	Electrolyte leakage for long cycling
PMMA	Hard chains	High ionic conductivity High affinity to electrolyte	Poor mechanical strength

Significantly, the solid electrolyte membrane framework provided an efficient approach to address the challenge of interfacial contact. Chen et al. introduced the concept of a cathode-supported solid electrolyte membrane,¹⁴⁴ which was obtained by directly casting the solid electrolyte on the cathode layer, as shown in Figure 18E. The SEM image in Figure 18F shows the close connection between the electrolyte and cathode in the case of indistinguishable interfacial layer. The improvement in the interfacial contact resulted in larger capacities and reduced the overall impedance. Moreover, a superior rate capability could be obtained.

Three-in-one batteries

Inspired by the battery construction design, membrane materials are developed in integrating three functional units (cathode, interlayer, and separator) into an efficient composite (Figure 19A,B),¹⁵⁷ ensuring a high-flux, flexible, high conductivity, and excellent rate performance in lithium metal-based batteries. Electrospun fibrous membranes help to eliminate the negative interface effects and show a high flexibility without the detachment of active material and widespread cracking after flexing only 10 times, as shown in Figure 19C,D. Figure 19E–G shows the morphology of integrated flexible S-CNTs/CoNCNFs/PVDF membranes, besides the advantage of stackable feature, the binary hierarchical architecture with embedded ZIF-derived carbon flakes also drove synergistic boosts to polysulfide confinement, electron transfer, and lithium-ion diffusion. It is apparent that both areal capacities and specific capacities of three-in-one batteries used in this work are higher than other similarly reported materials in Li-S batteries. Hence, the membrane materials provide a versatile and effective strategy for preparing integrated flexible membranes based on the cell construction engineering, which have practical significance for the future development in high-energy wearable and portable storage systems.

FIGURE 19. (A) Schematic illustrations of the conventional Li-S battery, (B) integrated flexible Li-S battery system; (C) A conventional Al foil electrode before and after bending 10 times, (D) the integrated flexible S-CNTs/CoNCNFs/PVDF membrane before and after bending 100 times; (E) SEM image of S-CNTs; (F, G) SEM images of CoNCNFs layer; (H) comparison of areal capacities of S-CNTs/CoNCNFs/PVDF membrane and other reported electrospun-based membrane materials, (I) comparison of specific discharge capacities in this study and other similar reported materials157

PROSPECTS

The applications of membranes in LRBs have recently received more attention because they provide an efficient strategy to obtain high-energy densities, serving as traditional separators or novel configurations in the forms of electrodes and electrolytes. Different fabrication methods have their respective features. A clear and complete understanding of the relationship among the preparation, structure, and performance is essential for designing the membrane structure and optimizing the chemical composition of the membrane materials. The requirements for the porous membranes employed in different configurations of batteries are diverse, including good mechanical strength, high thermal stability, fast charge transfer, easily electrolyte take-up, and excellent ionic conductivity. By optimizing the preparation parameters, such as functional coating and incorporation of inorganic particles and nanomaterials, the appropriately modified membranes can be employed to assist performance improvement.

With the improved ion/electrolyte transportation and charge transfer properties as electrodes/separators/interlayers, the membrane technology can serve as the component of LRBs and solve the vital issues faced by the emerging battery requirement. However, there still remains the challenges for the advanced energy storage technology, such as how to boost the gravimetric and volumetric energy density of batteries via rationally designing the porous structure and thickness of the membrane, how to further promote electron transport within the membrane electrode and to facilitate electrolyte accessibility with high loading of active materials through appropriate engineering design and optimization, as well as theoretical research which can provide microscopic insights of the electrochemical reaction kinetics and guide design directions of the battery architectures. The fundamental mechanisms, including the electrochemical reaction, self-consistent field theory, pore size exclusion, competitive effect, molecular dynamics simulations, as well as the precise control of the membrane structure, should be further explored for significant breakthroughs and achievements. Moreover, the development of advanced battery characterization methods and tools can greatly contribute to the fundamental mechanisms, such as capturing the dynamic reaction process and in situ observations of mass transport behavior in electrodes. Therefore, for future studies, experiments combined with theoretical investigations should be emphasized to accelerate their further improvement in emerging LRBs. Certainly, many researches have contributed on the architecture optimization for a longer and cycle life as well as a higher energy density. Benefited from the peculiarity of structure and composition, the membrane technology plays an important role in the novel battery constitution framework. What's more, integrated battery design is also a promising method to obtain the prominent transfer kinetics and robust mechanical properties.

With an increasing demand for flexible energy storage devices, the membrane technology in LRBs may offer a promising strategy for stretchable and wearable electronic devices. An economical, efficient, and large-scale membrane technology can complement the successful commercialization of LRBs.

ACKNOWLEDGMENTS

G.Y. acknowledges the funding support from the Welch Foundation F-1861 and Camille Dreyfus Teacher-Scholar Award. X.L. and G.H. acknowledge the funding support from Natural Science Foundation of China (21676043, 21506028); National Key Research and Development Program of China (2019YFE0119200); Liaoning Natural Science Foundation (2021-MS-116); Science Fund for Creative Research Groups of the National Natural Science Foundation of China (22021005); Dalian Innovation Funding supporting (2019J12SN68), the Liaoning Revitalization Talents Program (XLYC2007040), the Fundamental Research Funds for the Central Universities (DUT21YG113).

CONFLICT OF INTEREST

The authors declare no conflict of interest.